can also inactivate membrane and cytosolic enzymes. For example, ATPase and some glycolytic enzymes are inhibited by ethanol in a non-competitive manner (Banat et al., 1998; Ding et al., 2009; Ingram, 1976). Alcohols with longer carbon chains have further negative effects. Butanol can insert into the cell membrane and break hydrogen bonds between lipid tails increasing its toxicity as compared to ethanol (Ly and Longo, 2004).
Some possibilities to ease toxic effects of alcohols on producing strains include in-line product extraction to detoxify alcohols of the environment where the cultures grow, or the continuous recycling of the cells culturing the bacteria in a fresh media (Gaddy et al., 2007; Phillips, Klasson, Clausen, & Gaddy, 1993). However, these approaches have also a negative impact on the overall economics of the process. Apart from these, genetic modification of microbial strains has the potential to improve the alcohol tolerance, as well as, other metabolic features according to the desired needs (Tomas et al., 2003).
Butyribacterium methylotrophicum, butanol concentration was also among the highest value ever achieved (2.7 g/L) (Grethlein et al., 1991).
MBBR are characterized to retain biomass on an inert carrier supporting microbial growth and hindering biomass to flow out through the outlet (Hickey, 2009). The gas injection system in these reactors provides mixing by creating eddy currents in the fermentation broth. A study using an active culture of “C. ragsdalei” in a medium scale MBBR (36 m3) reported an ethanol concentration of 30 g/L after 30 days of operation with a productivity of 2.2 g/Lh (Hickey, 2009).
Gas in BCRs is sparged in the bulk liquid and no mechanical agitation is needed.
BCRs may allow satisfactory heat and mass transfer rates (Abubackar et al., 2011), although these may be limited by coalescence or back mixing of gaseous components, even at high flow rates (Acharya et al., 2015; Munasinghe and Khanal, 2010b). This technology was primarily developed for industrial use with large working volumes.
Alcohol production by C. carboxidivorans P7 was observed in a 4 L BCR in a batch mode with a producer gas at production rates lower than 0.042 g/Lh, obtaining concentrations of 5.00 and 0.75 g/L of ethanol and butanol, respectively (Datar et al., 2004).
Trickling filters (TF) characterized by a continuous feed of a liquid trickling phase and the use of inert packing materials are commonly used for different gas treatments (Kennes et al., 2009). The main parameters influencing the mass transfer in TF are reac- tor packing and material size, liquid recirculation rate, and gas flow rate (Bredwell et al., 1999). In TF, plug flow is easily achieved and there is no need of mechanical agitation (Ahmed and Lewis, 2007). A study comparing the performance of continuous STR, BCR, and TF for syngas fermentation concluded that higher CO conversion rates (>80%) and higher productivities were achieved in a TF (Klasson et al., 1992).
In MBRs, microbial cells are attached to the membrane surface in the form of a biofilm. Syngas is injected into the hollow fiber and the process gas flows across the membrane toward the biofilm, enhancing the mass transfer. Different kinds of MBR
containing microporous layer membranes to support a biofilm on the outer surface have been described (Datta et al., 2009; Hickey et al., 2011; Tsai et al., 2011). One of the major disadvantages of this system is that the liquid may enter the pores owing to variation in pressure across the membrane, leading to a phenomenon known as pore-wetting. The use of microporous membranes with a biofilm of “C. ragsdalei” produced an ethanol concentration of 10 g/L after a 20 days continuous operation with an ethanol productivity of 0.1 g/Lh (Hickey et al., 2011). The productivity of ethanol by the same species was increased to 0.28 g/Lh, reaching ethanol concentrations of 13.3 g/L, when a sandwiched type combination of membranes having a liquid impermeable layer between two microporous membranes was used (Datta et al., 2009). Some other MBRs use hydrophilic membranes instead to separate the gas to the liquid phase. By using this approach “C. ragsdalei” presented the same productivity as in microporous membrane (0.05 g/Lh) (Hickey et al., 2010). Although the use of hydrophilic asymmetric membranes has also been tested for the ethanol production (4.2 g/L), its productivity reached values of 0.08 g/Lh using “C. ragsdalei” (Datta et al., 2010).
(i) (ii) (iii) (iv)
(v) (vi)
Figure 1.3 Schematic representations of STR (i), MBBR (ii), BCR (iii), TF (iv), and MBR (v-vi). Gas (1) and nutrient (2) feed into the reactor, liquid products from the reactor (3), gas outlet (4), agitator (5), carrier retainer (6), biomass retainer (7), sparger (8), recirculation (9), pump (10), packed bed (11), nutrient recirculation (12), and membrane modules (13). A: cross section of a microporous membrane present in modular membrane supported bioreactor and nutrient recirculation, B: cross section of microporous membrane present in membrane supported bioreactor, C: cross section of a hydrophilic membrane having biofilm growth on the membrane surface, D: cross section of an asymmetric membrane present in stacked array bioreactor and horizontal array bioreactor. a: microporous membrane, b: hydrophilic membrane, c: microorganism (biofilm), d: liquid phase, e: medium inlet and liquid products from the membrane, f: gas inlet to the membrane, g: liquid impermeable layer, h: biolayer, i: hydration layer. Adapted from Abubackar et al (Abubackar et al., 2011).
e 1.2 Ethanol and butanol production using different bioreactors, bacteria, volume, and ethanol and butanol productivity and concentration using syngas fermentation. References (Phillips et al., 1993) (Mohammadi et al., 2012) (Gaddy et al., 2007) (Liu et al., 2014) (Grethlein et al., 1991) (Hurst and Lewis, 2010) (Kundiyana et al., 2010b) (Maddipati et al., 2011) (Chang et al., 2001) (Rajagopalan et al., 2002) (Datar et al., 2004) (Richter et al., 2013) (Shen et al., 2014) (Tsai et al., 2012) (Hickey et al., 2010) (Hickey et al., 2011) (Datta et al., 2009) (Datta et al., 2010) (Hickey, 2009) Abbreviations: STR: Stirred tank bioreactor; BCR: Bubble column reactor; MBR: Membrane bioreactor; SMSR: submerged membrane supported bioreactors; MSB: membrane supported bioreactor; SAB: Membrane modules in axially stacked array bioreactor (MBR); MBBR: Moving bed biofilm reactor. a: hollow fiber membrane; b: hydrophilic membrane; na: not applicable; nr: not reported
(g/L) Butanol (na) (na) (na) (1.0) (2.7) (nr) (0.47) (0.00) (0.00) 1.10 0.75 na <0.45 nr nr nr nr nr nr
Ethanol 48.00 6.50 26.00 8.0 0.33 2.00 25.26 9.60 0.09 1.30 5.00 21.00 23.93 15.00 10.00 10.00 13.30 4.20 30.0
Productivity (g/Lh) Butanol na na na 0.017 0.041 na nr 0.027 na 0.012 0.006 na nr na na na na na na
Ethanol 3.570 0.049 0.650 0.088 0.005 0.020 0.036 0.027 0.950 0.014 0.042 0.370 0.120 0.125 0.050 0.100 0.277 0.084 2.2
Culture volume 1 2.0 2.5 3.3 1.2 0.35 70.0 3.0 0.2 4.5 4.0 1.0 8.0 3.0 3.0 2.0 nr 3.0 1.8·104
Microorganisms Clostridium ljungdahlii Clostridium ljungdahlii Clostridium ljungdahlii Mix culture Butyribacterium methylotrophicum Clostridium carboxidivorans “Clostridium ragsdalei” “Clostridium ragsdalei” Eubacterium limosum Clostridium carboxidivorans Clostridium carboxidivorans Clostridium ljungdahlii Clostridium carboxidivorans “Clostridium ragsdalei” “Clostridium ragsdalei” “Clostridium ragsdalei” “Clostridium ragsdalei” “Clostridium ragsdalei” “Clostridium ragsdalei”
Bioreactor characteristics Quimiostate Quimiostate Quimiostate Quimiostate Quimiostate Batch Batch Batch Continuous Batch (first 96 h) Batch with gas producer Batch STR for 2 days switched to BCR Continuous Continuous Batch mode for 2 days switched to continuous Continuous Continuous Batch mode for 46.5 h switched to continuous Continuous
Bioreactor STR STR STR STR STR STR STR STR BCR BCR BCR STR and BCR MBRa MBRa SMSR MSBb MSBa SAB MBBR
The development of new bioreactor designs for syngas fermentation and alcohol pro- duction was out of the scope of the present PhD dissertation, and has been covered here just briefly. We found it important to review how the use of different bioreactors and operation configurations has been targeted as a research objective to increase alcohol productivity. However, the work in this thesis focuses mainly to study some of key physiological properties of model carboxydotorphic bacteria and anaerobic hungate tubes and bottles, all operated in batch conditions, have been chosen for all experiments.
The aim of this PhD was to provide a deeper insight into key physiological aspects of syngas fermentation by carboxydotrophic clostridia, with special focus on clarifying key factors about molecular mechanisms that enhance alcohol production.
Specific objectives:
1. To understand the impact of added formate on kinetic parameters for growth and alcohol production. The reducing power demand for formate dehydrogenase activity in WLP has the potential to diminish due to the assimilation of formate as a partially reduced carbon source.
2. To test the effect on growth rate and alcohol production of pH, temperature, and organic matter supplements as key factors capable of controlling growth and shift from acidogenesis to solventogenesis in carboxydotrophic species.
3. To determine inhibitory effects of ethanol and butanol during syngas fermenta- tion. Alcohol toxicity is not known in detail and it could be useful especially in fermentation modeling using continuous reactors.
4. To discover new isolates with the potential to transform inorganic carbon to ethanol or butanol as biofuel in order to increase the small number of syngas-fer- menting bacterial strains known to date and, may be also, the range of operational conditions in which bio-ethanol and bio-butanol can be produced from inorganic carbon.
3.1 Bacterial strains.
All experiments were conducted using Clostridium ljungdahlii PETC (DSM13528T), Clostridium carboxidivorans P7 (DSM15243T), and/or Butyribacterium methylotrophicum DSM3468. The three strains were obtained from the DSMZ culture collection as a freeze dried cultures.
3.2 Media and culture conditions.
Bacteria were cultured in an anaerobic mineral medium modified from ATCC 1754 PETC (Table 3.1). Differences from ATCC 1754 PETC were: (1) all soluble carbon sources, i.e. yeast extract, fructose, and NaHCO3 were excluded from the original formulation; (ii) 2-(N-morpholino)ethanesulfonic acid (MES) (100 mM, final concentration) was used instead of bicarbonate as pH buffer.
Liquid medium was prepared with the mineral medium, Wolfe’s vitamins solution, trace elements solution, and resazurin. The pH of the medium was initially adjusted to 6.0 with 1M NaOH. After that, the medium was boiled to remove dissolved oxygen and distributed anaerobically in butyl rubber 125 mL glass bottles (for culture maintenance) and 25 mL Hungate tubes (for fermentation experiments). Tubes and bottles were flushed with N2 and sealed before sterilization in the autoclave. MES buffer and reducing agents were added to the autoclaved medium from sterile anaerobic stock solutions.
Headspace was flushed with synthetic syngas (CO:H2:N2:CO2 [32:32:28:8]) of high purity (Praxair Technology Ltd, Spain). All culture manipulations and inoculation of freshly prepared media were done in an anaerobic chamber (gas mixture N2:H2:CO2
[90:5:5], Coy Lab Products, Michigan, USA).
Cultures were incubated in gas-tight closed bottles outside the anaerobic chamber in a Stuart incubator SI500 (Bibby Scientific Limited, OSA, UK), and were maintained ac- tive by weekly transfers of 1 mL of culture (1:25 ratio) into serum bottles containing
25 mL of modified ATCC 1754 PETC medium (Figure 3.1A). The headspace of bottles was filled with syngas at an overpressure of 100 kPa at the beginning of the incubation period.
Table 3.1. Composition of the mineral medium, reducing agents, trace elements solution, and the Wolfe’s vitamin solution for the preparation of modified ATCC 1754 PETC medium.
Mineral medium Reducing agents
Element Final concentration
(g/L) Element Final concentration
(g/L)
NH4Cl 1.0 L-Cysteine·HCl 0.4
KCl 0.1 Na2S·9H2O 0.4
MgSO4·7H2O 0.2
NaCl 0.8
KH2PO4 0.1
CaCl2·2H2O 0.02
MES 19.52
Resazurin 0.001
Wolfe’s Vitamin Solution Trace elements solution Element Final concentration
(mg/L) Element Final concentration
(mg/L)
Biotin 0.02 Nitrilotriacetic acid 20
Folic acid 0.02 MnSO4·H2O 10
Pyridoxine-HCl 0.10 Fe(SO4)2(NH4)2·6H2O 8.0
Thiamine-HCl·2H2O 0.05 CoCl2·6H2O 2.0
Riboflavin 0.05 ZnSO4·7H2O 0.002
Nicotinic acid 0.05 CuCl2·2H2O 0.2
D-Ca-pantothenate 0.05 NiCl2·6H2O 0.2
Vitamin B12 0.001 Na2MoO4·2H2O 0.2
p-Aminobenzoic acid 0.05 Na2SeO4 0.2
Thioctic acid 0.05 Na2WO4 0.2
3.3 Fermentation experiments.
Cultures in the exponential growth phase were used as inoculum for batch experi- ments. Fermentation experiments were conducted in 25 mL Hungate tubes containing 6 mL of modified ATCC 1754 PETC medium (Figure 3.1B). A 10% inoculum was used in all experiments. Culture tubes were thoroughly flushed with syngas mixture for at least 1 min. Headspace overpressure was adjusted to 100 kPa. Depending on the experiment syngas was injected only at the beginning of the experiment, or once a day to ensure re- plenishment of gas substrates. The incubation was done under mild agitation (100 rpm) in a Stuart incubator SI500 shaker. Tubes were incubated horizontally to enhance gas-
liquid mass transfer. The same parent culture was used to inoculate all tubes used in a single kinetic experiment to prevent possible biases due to differences in the inoculum.
Growth was monitored on a daily basis by measuring the absorbance at 600 nm using a spectrophotometer. Samples for the determination of organic acids (formate, acetate, butyrate, and caproate) and alcohols (ethanol, butanol, and hexanol) concentrations were obtained regularly from the liquid media. Liquid samples were filtered using nylon filters (0.2 µm pore size) to remove cells, and stored at 4 ºC until analyzed. Headspace samples for the analysis of gaseous substrates (CO, CO2, and H2) were also collected with a syringe and distributed in BD Vacutainer tubes. Vacutainer tubes (Sangüesa SA, Spain) were stored at room temperature. The pH of the media was measured using a BASIC 20 pH meter.
3.4 Determination of growth variables
Linear regression of natural logarithm transformed absorbance readings (lnAt) at time intervals (t) were used to estimate growth rate, according to equation (Eq. (3)):
lnAt= lnAinit+ μ · t (3)
Growth rates (µ, h-1) were calculated for each incubation experiment as a measure of the growth capacity of the bacterial cultures at the conditions set in the different ex-
Figure 3.1. A: Image of 125 mL serum bottles used to maintain cultures of the three carboxydotrophic bacteria in active state. B: Fermentation experiments conducted in 25 mL Hungate tubes.
A B
periments. Duration of the lag phase was estimated as the time interval between the in- oculation of tubes and the time at which the maximum growth rate was observed.
3.5 Analytical methods
3.5.1 Protein concentration
Concentration of proteins was determined following a protocol adapted from Scott et al. (2011). Frozen pellets of 1 mL bacterial culture were suspended in 1 mL of 40 mM Tris at pH 7.0. Cells were disrupted by sonication. Six sonication cycles, each cycle including 30 s of sonication at maximum intensity and one minute on ice, were performed using a tip probe (5 mm) sonicator (B.Braun Labsonic 2000). Cellular debris were removed by centrifugation at 16000 rcf for 30 min. Protein quantification was performed using a QubitTM kit (Qubit® 2.0 Fluorometer, Invitrogen) and the recommended method.
3.5.2 Gas composition
Composition of CO2, CO, H2 and N2 (% vol) in gas samples was analyzed using a gas chromatograph (Agilent 7890A GC system, Agilent Technologies, Spain) equipped with a fused zeolite capillary column (HP-Molesieve, 30 m x 0.53 mm x 50 µm) and thermal conductivity detector. Helium was used as the carrier gas. The injector and detector temperatures were set to 115 ºC and 275 ºC, respectively. The oven temperature was initially kept at 45 ºC for 6 min, and subsequently increased following a ramp of 8 ºC·min-1 until 70 ºC, holding the temperature 2 min, 5 ºC·min-1 from 70 ºC to 130 ºC and 35 ºC·min-1 from 130 ºC to 220 ºC. The cycle finished at temperature of 220 ºC for 5 min.
3.5.3 Organic acids and alcohols
Concentrations of organic acids and alcohols were determined using an Agilent 7890A GC system (Agilent Technologies, Spain) equipped with a fused-silica capillary column (DB-FFAP, 30 m x 0.32 mm x 0.5 µm) and a flame ionization detector. The injector and detector temperatures were set to 250 ºC and 275 ºC, respectively. The oven temperature
was initially kept at 40 ºC for 1 min, and subsequently increased following a ramp of 5 ºC·min-1 until 70 ºC, 10 ºC·min-1 from 70 ºC to 180 ºC and 35 ºC·min-1 from 180 ºC to 250 ºC. Finally, temperature was kept at 250 ºC for 5 min.
3.6 Calculations
Concentrations of undissociated acids (UA) were estimated based on actual pH of media and the total concentrations of acid (the salt and the acid form) according to the equilibrium equation (Eq. (4)).
UA = Ct – (10(pH-pKa) · Ct)/( 10(pH-pKa) + 1) (4)
Where Ct is the concentration of acid measured in the GC and pKa is the logarithmic of the acid dissociation constant. The formate pKa value used was 3.76 corresponding to the equilibrium constant at 35 ºC (Kim et al., 1996). The pKa values for acetic acid were 4.76 and 4.77 at 25 ºC and 37 ºC, respectively, pKa for butyric acid were 4.82 and 4.88 at corresponding temperatures and pKa for caproic acid were 4.88 at both temperatures (Creager and Clarke, 1994; Harned and Ehlers, 1933; Zigová et al., 1999).
3.7 Statistical analyses
All statistical analyses were conducted using SPSS 15.0 statistical package for Windows (LEAD Technologies Inc., EEUU). Significance levels were established for p≤0.05. ANOVA test were used to analyze differences of growth parameters or alcohols and acid production to initial different variable among treatments. When multiple comparisons were required, a T3 of Dunnet post-hoc test or Bonferroni post-hoc test were used assuming non equal variances or equal variances between treatments, respectively. Kruskal-Wallis test was used when non-parametric tests were required.
Pearson correlation tests were used to analyze possible correlations between experimental data.
The results of this PhD dissertation have been organized in six different chapters outlined in the form of scientific contributions to research journals. Most of these chapters have already been published or had been submitted to ISI journals at the time this dissertation was written.
4.1
Clostridium ljungdahlii PETC and Clostridium
carboxidivorans P7 grown on syngas
Part of this chapter has been published as:
Ramió-Pujol, S.; Ganigué, R.; Bañeras, L.; Colprim, J., 2014. Impact of formate on the growth and productivity of Clostridium ljungdahlii PETC and Clostridium carboxidivorans P7 grown on syngas.
International Microbiology 17: 195-204. DOI: 10.2436/20.1501.01.222
4.1.1 Background
The first steps in the methyl and carbonyl branches of the WLP, formate synthesis and the carbon monoxide formation, are recognized as reducing equivalent sinks that may diminish the growth capacity of cells and the conversion of inorganic carbon into valuable chemicals (Daniell et al., 2012). The use of hydrogen and carbon monoxide as sources of reducing equivalents is maximized during autotrophic growth. In this condition, the assimilation of formate as a partially reduced carbon source has the potential to reduce the hydrogen/CO demand for formate dehydrogenase activity in the WLP (Brown et al., 2014; Schuchmann and Müller, 2013). We hypothesize that the excess hydrogen can further be diverted to acetate reduction, increasing biofuel production. The synthesis of formic acid from CO2 has been accomplished in a bioelectrochemical system (BES) using a purified formate dehydrogenase enzyme (Srikanth et al., 2014). Electrosynthesis coupled to fermentation by carboxydotrophic bacteria, has been proven but not studied in detail (Nevin et al., 2010). However, the concept of enzymatic electrocatalysis involving energy applications is gaining in prominence, especially in the direction of enzymatic electrosynthesis of desired chemicals and fuels under non-limiting reducing power supply.
Formate has been reported as an inducer of acetate production in Clostridium acetobutylicum. Maximum effects of formate on acetate production in C. acetobutylicum were obtained under acidic conditions (at pH=4.8) (Ballongue et al., 1985). Despite this exam- ple, growth on weak organic acids is rather difficult for most microorganisms and inhibition occurs at very low concentrations. Inhibition effects are higher at low pH values where higher concentrations of the undissociated acid forms exist, which can freely diffuse to the cytoplasm of the cell eventually causing the dissipation of energy gradients built across the cell membrane (Jones and Woods, 1986). Additionally, formate can cause sub-lethal damage in some bacteria and has been used as an antibacterial agent (Thompson and Hinton, 1997).
PETC has been grown chemoorganotrophically in a medium containing 5 g/L of formate and 1 g/L of yeast extract (Tanner et al., 1993). However, similar experiments have never been done autotrophically with this strain. Moreover, B. methylotrophicum can also use formate as substrate for growth, but it is unclear whether other acetogenic bacteria, including P7, can use formate when growing either organo- or autotrophically (Kerby and Zeikus, 1987; Liou et al., 2005). In this light, the present chapter hypothesizes that the addition of formate, as a partially reduced C1 compound, would positively impact kinetic parameters for growth and alcohol production in C. ljungdahlii PETC and C. carboxidivorans P7 by diminishing the need for external reducing equivalents. The aim of this chapter was to provide experimental evidence to evaluate formate addition as a potential enhancer of alcohol production in C. ljungdahlii PETC and C. carboxidivorans P7.
4.1.2 Experimental design
4.1.2.1 Fermentation experiments.
Exponentially growing C. ljungdahlii PETC and C. carboxidivorans P7 cultures were used as inocula for batch experiments to test for formate effects on growth and alcohol production. Syngas was injected only at the beginning of the experiment. Sodium formate solutions adjusted at the desired pH (5.0, 6.0 or 7.0) were aseptically added to the medium at final concentrations of: 0.1, 1.0, 2.2, 5.5, 7.6, 10.9, 15.0, 20.0, 27.2, 54.5, and 109.0 mM. In all batch tests, tubes containing no sodium formate were included as controls for growth kinetics under fully autotrophic conditions. Experiments were carried out for the two bacterial species at three pH values, 5.0, 6.0, and 7.0. The cultures were incubated at 35 ºC under mild agitation on a rotary shaker. All experimental conditions were assayed in triplicate using three independent inoculated cultures.
Growth experiments finished once cultures reached the stationary growth phase, which was considered to occur 48 to 72 h after growth cessation. Samples for the deter- mination of organic acids (formate, acetate, and butyrate) and alcohols (ethanol and